AI4TRUST is a European initiative involving 15 consortium partners and 2 associated partners from 11 countries. The Horizon Europe Programme of the European Union is funding this EUR 6 million project to use AI to fight disinformation.

Fake news has always existed. But today it is disseminated faster and further than ever before. This exacerbates crises, deepens divisions, incites political and physical harm, damaging the effective functioning of democratic societies in Europe. Visual imagery and speech can be re-framed to mislead viewers about events.

The synthetic generation of disinformation using AI speeds up manipulated content development and dissemination, making it harder to mitigate its effects. By 2026, close to 90% of Internet content may be artificially created.

A coordinated effort from EU countries, EU institutions, online platforms, news media, and citizens is essential to address the significant issue posed by large- scale disinformation campaigns in Europe. Everyone has a role to play in ensuring information flows are trustworthy so we can share news as responsibly as possible.

Monitoring and analysing large volumes of online data is beyond human ability. But we can use AI to protect the integrity of information we share and limit the spread of disinformation online. Technological solutions can complement other measures such as the implementation of the strengthened Code of Practice on Disinformation.

Over the next three years, the AI4TRUST consortium will develop and deploy hybrid solutions using AI to enhance human-centred responses. Innovative algorithms will equip researchers and media practitioners with a new toolbox.

AI4TRUST will develop innovative analytical solutions using AI to filter massive data streams and identify disinformation related to topics subject to mid-to-large scale manipulation, such as information on health-related issues, climate change and further social challenges.

AI can handle complex data streams across languages and borders, to organise processed multimodal data such as online text or video content into smaller, semantically consistent information packages, so human fact-checkers can work more efficiently.

AI can detect inconsistencies and misrepresented information, as well as its estimated reach on social media.

This allows experts to respond faster, with better counter-arguments and ways of re-contextualising images and data. This should make it easier to raise public awareness and reduce incentives for individuals to create and circulate fake news.

AI4TRUST will establish a common Open Observatory platform to give media practitioners, policy makers and research scientists key, nearly real-time elements to assess and counter disinformation.

They will be able to carry out personalised queries and analyses within a trusted environment. AI4TRUST will leverage current state of the art solutions to develop innovative digital tools built on a foundation of ethical values. This will embed European values in the way we communicate and share information in Europe.